English
Embedded World 2025 brought together some of the world’s leading tech experts and allowed companies to show off their latest hardware and software solutions. Developers, technical managers, and engineers from around the globe attended the Nuremberg, Germany-based event.
Three products by Jumptec, Versalogic, and Connect Tech stood out to the All About Circuits team for their innovation as AI-accelerated embedded computing solutions. All three of these products target edge AI and industrial manufacturing markets.
Jumptec SMARC Card Sports TI AM67 Jacinto Processor
The new SMARC-sAM67 board from Jumptec leverages Texas Instruments' dual/quad AM67x Jacinto processor. The SMARC-sAM67 provides the computing power to perform 3D rendering at high resolution, making it useful for robotics applications that use machine vision or multi-camera systems. The rendering abilities also mean robotic teaching pendants are quickly integrated into the model, speeding up robot deployment times.
The SMARC-sAM67.
The SMARC-sAM67 also boasts low latency and fast access to the 32 GB of e.MMC storage and the 8 GB of LPDDR4 RAM. It also includes dual 10 Gigabit Ethernet ports, 3.0 USB, and PCIe Gen 3 for versatile connectivity options.
Versalogic Intros Ruggedized Edge Computer
Versalogic designed the Sabertooth AI Connect , a ruggedized edge computer, for the harshest of industrial environments. It can withstand temperature extremes of -40℃ to 85℃ (-40℉ to 185℉) and meets MIL-STD-202H specifications for shock and vibration. The broad temperature range and resistance to vibration make it a good fit for aerospace, defense, energy, medical, and other critical industries. Versalogic says that the Sabertooth AI Connect, like its other devices, is desgind for long term availabilty, typically a 10 year product lifecycle.
Sabertooth AI Connect packs AI-enhanced computing into a compact package.
In terms of computing power, the Sabertooth AI is powered by the hex-core Intel Xeon-E CPU and the Nvidia RTX 2000 ADA GPU. The combination of these processors enables generative AI, including 3D rendering and high-performance interfacing. More specifically, the Ada Lovelace architecture found in the GPU supports 3072 CUDA cores, 96 Tensor cores, and 24 RT cores. All of this processing capability means that the floating point performance can be up to 12.99 TFLOPs.
The edge computer is equipped with dual, independent, 10 Gbit Ethernet ports as well as three mini DP ports for video streaming at 8k.
Connect Tech Miniaturizes SOM Platform
The Essential EdgeAI Network Node combines the Nvidia Jetson system-on-module (SOM) platform with a small physical footprint. Designed for edge computing in industrial settings, these nodes are powered by 12 V to 17.5 V DC and communicate via Gigabit Ethernet connection, UART, or USB 3.2. They operate “headless,” meaning there is no display, further reducing the space requirements.
Connect Tech Essential EdgeAI Network Node.
These network nodes are designed to handle generative AI applications, including images and video processing. The platform supports machine vision and machine learning algorithm development and control responses, not simply pattern recognition.
Priced for high volumes and backed with the Connect Tech Board Support Package (BSP), the Essential EdgeAI Network Nodes are fully supported from the start of installation through its full deployment.
Hardware Leads the Future of Edge Computing
The edge computing market will continue to grow and drive new innovations in the industrial and manufacturing market segments. As industrial facilities recognize how to collect and process data at their source location, leveraging AI to increase the speed of control decisions will become vital. Jumptec, Versalogic, and Connect Tech provide examples of how hardware solutions are leading this next wave of edge computing and AI applications.
